- Despite marginal improvements, the municipal bond market remains inscrutable to many.
- At the top of most gripe lists is investment cost; just what are reasonable mark-ups?
- This article offers an expert's insight.
The difference between what a muni dealer pays for a bond in the open market and the price at which it might sell that bond to a customer is known as a mark-up. There's been a lot of talk about muni mark-ups of late, notably in a Wall Street Journal article from March 10.
Dealers say that they are entitled to mark-ups that fairly compensate them for the risks and costs they assume in furnishing liquidity for investors. Some academics say that dealers are overcharging, while retail investors have very little idea about what's going on, but are feeling somehow exploited. The reality is that they all have a point, but some simple calculations show a high probability that current market practice does need to shift toward more favorable treatment of investors.
Historically, retail muni spreads have ranged from about $5 per bond ($1,000 par value) for shorter-term bonds (up to about 10 years to maturity) to as much as $30 per $1,000 for long term bonds (20 years or more.) The current average is estimated to be about $17 per bond. But what goes into that average?
It is commonly held that some larger brokerage firms charge more than many smaller firms. This is in part because the larger firms have more significant cost components to their trading and distribution operations, as follows:
1. Market Risk -- Since 2008, most fixed-income markets have experienced diminished liquidity, while some firms have altogether left the market as principals. Other large firms have combined, leaving fewer large firms to furnish liquidity -- i.e., a willingness to buy bonds from customers without an offsetting purchase order from another buyer. Given the recent negative credit trend among municipalities, this means increased loss exposure over time.
2. Research -- The muni market has always been more obtuse than other U.S. credit markets for the simple reason that municipalities are not forced to file the kind of information and disclosures that are incumbent upon corporate issuers of debt. While corporate credit analysts can readily draw from available public data and borrow liberally from equity analytical work, muni analysts don't have these luxuries and must spend a lot of time extracting data from infrequent filings and from thinly staffed local government finance departments. Additionally, muni bond structures are now often more complex than in the past, so for these reasons it's simply more labor intensive to compile useful credit research for municipal bonds.
3. Scavenger Hunting -- Finding suitable municipal bonds that fit a specific individual criteria set for an investor is time-consuming. Sourcing a muni bond with the right rating quality, maturity, credit risk profile, and state tax exemption can be a bit of an Easter egg hunt and often smaller firms must rely on larger firms to carry those special finds in their inventories, as their own firms do not even maintain inventory (see Figure 2 below).
4. Economies of Scale: Measured by trading volume, the muni market is but a fraction of the corporate debt market, implying fewer trades at the same basic unit cost for dealers (e.g., trade processing, clearing, reporting, etc.; see Figure 1). Moreover, nominal demand volume for taxable corporate debt is far greater than for munis, as there are many more and larger buyers of taxable debt, like pensions, banks, central governments, insurance companies, etc. Most of the demand for munis comes from individual investors, and that buyer group relies on the dealer community for research and other market information to a much greater extent than do more sophisticated buyers. In addition to this, turnover is lower for munis as tax-exempt bond investors tend to hold them longer than buyer classes of corporate debt.
5. Smaller muni dealers are today commonly forced to rely on the muni bond inventories of larger firms when filling orders for their own customers, more so than prior to 2008. This implies a much greater frequency of added mark-ups between the supplier dealer and the smaller firm serving its customer, but the middle-man dealer must take a small mark-up if it is to even have a chance competing for business against larger shops. These inter-dealer trades are tracked by the Municipal Securities Rulemaking Board, and because of the growing dependence of some dealers on larger dealers' inventory, interdealer trades have grown as a percentage of all customer trades. In 2004, interdealer trades were 12.7% of all customer buys and sells of munis. In 2012, that ratio was 24.3% (source: MSRB Fact Books, 2008 through 2012).
6. Oligopoly: Today, a larger proportion of trading volume is handled by fewer and larger firms. In 2012, the top 10 firms handled 72.4% of all customer trades in municipal bonds, and that's out of a total of 1,645 registered dealers. This is due to firm consolidation. Morgan Stanley is comprised of former bulge-bracket muni firm Smith Barney, long a muni market behemoth; Bank of America is now comprised of Merrill Lynch; and Wells Fargo is made up of former muni shops Wachovia, Wheat First, and Butcher & Singer. That has resulted in a heavy concentration of secondary muni market trading in fewer organizations, all of whom would likely claim to have higher-than-the-industry-average overhead associated with muni business.
All that said, is an average spread of $17 a bond fair to retail investors? Let's look at the math, or at least some of the math.
The Wall Street Journal article referenced above said that retail investors traded $951 billion in munis from 2009 to 2014 (based on trades of $100 thousand or less.) If dealers earned a spread of $17 on each trade that averaged $100 thousand par value, that means they made $1,700 on each trade. That strikes me as a per-trade number that would rather wildly exceed an average ticket cost, even if you factor in research, trade finance, market risk, processing, clearing, salaries, etc. Charles Schwab does not have most of the overhead costs that a large muni dealer does, but that notwithstanding, they only charge $125 for a broker-assisted muni trade of $100 thousand, or less than 8% of what appears to be the average mark-up. I would very conservatively estimate that a large muni shop has per-trade costs somewhere between $300 and $500, so is a margin of 240% to 466% fair to investors? Doesn't sound like it, but it might be more equitable to look at things from the dealer's market risk perspective.
Assume a dealer holds a 15-year muni bond paying 5% interest in inventory in hopes of selling it to a customer, after having bought it from another customer, and has as its cushion against market loss the $1,700 described above. Now, to be fair, some portion of that spread for the "house" will be paid to the client-facing broker in the form of a commission, so the spread as a cushion against market risk is actually somewhat lower. Assuming the dealer was including, say, $15 per bond out of the $17 spread as commission, it would appear that the dealer only has a $200 loss cushion -- but we're not done. The $15 commission is not all paid to the broker, as he or she actually receives what is known as a net commission based on what is called a "pay-out rate." The industry norm for this is 30%, so, in fact, the dealer only gives away $4.50 of the $17 dollar spread, leaving a real cushion of $1,250.
Usually bonds move out of inventory within a few days or maybe a couple of weeks, but let's assume this dealer doesn't find a buyer for 30 days. During that period, market rates for the bond in position can rise by as much as 24 basis points before the dealer loses money. How likely is this to happen? Well, not highly likely, but it's not a rare occurrence either. Using the 10-year Treasury note as proxy, from 2009 to 2014, this occurred 20% of the time.
Of course, there is also the chance that rates could fall, allowing the dealer to raise their offering price and make even more money. In our proxy case, prices rose in 30 day periods about 32% of the time from 2009 to 2014 enough to offset the bid/ask spread ($17 per bond), so it would appear that the odds were 3-to-2 against losing money to market risk over the last five years. While not a perfect benchmark for our hypothetical 15-year muni bond, yield movements on the 10-year T-note fairly well paralleled yield movement on munis, when compared to the Bond Buyer U.S. Municipal Bond Index, as shown in Figure 3. Interestingly, during the last 30 years, the above scenario would have experienced a net loss 22% of the time and a gain 21% of the time (source: U.S. Federal Reserve Bank, Bloomberg).
One of the challenges investors face in discerning mark-ups in muni-land is reinforced by the fact that nearly all trades take place with dealers acting in a principal capacity. This means that they take title to municipal bonds before they resell them to customers, rather than simply brokering the trade on behalf of the customer, much like a real estate broker does. The difference in prices paid and received by a principal constitutes the mark-up and because the dealers have first bought the bonds with their money, they may choose to characterize their position in the transaction flow as principal. But these circumstances can vary and one variance belies the principal characterization.
When a dealer has an order in hand from a customer and then buys bonds from another dealer in instant fulfillment of that customer order, it is called a "riskless trade." This is much more like the real estate broker in that the buyer is known and ready to pay for and acquire the property. Trades like this are effected by the thousands every day and, in fact, are even more common today because more dealer firms rely on the bond inventory of other firms. And this reality gives rise to a scenario in which investors can begin to crack the force-field of mark-up opacity.
When a customer approaches a brokerage firm to buy munis, they should specify that any fulfillment of their investment inquiry that results in the dealer entering into a riskless trade should be done as an "agency" trade. It is the customer's right to do so and if the dealer agrees to the request, they must disclose the commission amount on the transaction confirmation. While this won't throw open mark-up practices to the full light of day, it will give investors a chance to gauge just how much a broker is taking as compensation and may even cause the broker to lessen that compensation.